Random Walk

假设$x_t, t=0, 1,2,\cdots$为一个以0为反射壁的一维随机游走:

First, $x_0 = 0$, For $t \ge 0$, $x_{t+1} = x_t + 1$ with probablity 0.5, and $x_{t+1} = \max\{x_{t}-1, 0\}$ with probability 0.5.

Let $e_t$ denote the expectation of $x_t$:

$$e_t = E \left\{ x_t \right\}, t=1,2,3,\cdots$$

显然,$e_t$单调非减。如果$e_t$收敛,令:

$$e_{\infty} = \lim_{t \to +\infty} e_t$$
那么,一定存在$N$,使得 $\forall n \ge N$,

$$e_{\infty} - 0.25 < e_n \le e_{\infty}$$
令$P(N,M)$表示从第$N$步到第$N+M-1$步,$x_t$至少回到原点一次的概率。
根据一维随机游走的常返性:

$$\lim_{M \to +\infty} P(N,M) = 1$$
显然,$P(N,M)$关于M单调非减。所以对于$N$,存在$M$,使得

$$P(N,M) > 0.5$$
易知:

$$e_{N+M} > e_N + 0.5 * P(N,M) > e_{\infty}$$
与假设矛盾。$\blacksquare$

Equation Templates

虽然觉得hexo蛮好玩的,但是目前还没有足够的动力用hexo作为blog。这里就作为一个临时记点东西的地方吧,估计以后很少会在这里写东西。

hexo有个缺点是,每次修改一点都把所有文件重新generate一遍,然后要commit一堆上去…

Read More

Get it done

I tried hexo on Github Pages for fun,
but don’t know why I should use it.
I may not post on this site much.
Test equations:
For any $x > 0$, $y> 0$,
$$G(x) = \int_0^x e^{-\frac{x^2}{2}} \mathrm{d}x.$$

Other examples:

$$\begin{aligned} \dot{x} & = \sigma(y-x) \\ \dot{y} & = \rho x - y - xz \\ \dot{z} & = -\beta z + xy \end{aligned}$$

The local website updates right after saving the md file, that’s nice.
After installing hexo-livereload plugin, the browser will automatically refresh as well!